We study the Sec7 domain guanine nucleotide exchange factors (GEFs) for the Arf family of small GTPases. We are interested in the roles of these proteins in membrane dynamics and protein trafficking. The Arfs and the Arf GEFs are important regulators of both organelle structure and protein transport throughout the cell. We are focusing our attention on the large Golgi-localized Arf GEFs involved in transport through the secretory pathway in both budding yeast and mammalian cells. A central question in cell biology is how the elaborate and dynamic structures of membrane systems are maintained in the face of constant trafficking into and out of each organelle. Particularly important questions include the way organelle structure is generated and maintained and how structure is correlated with the underlying molecular events of protein sorting and membrane remodeling. An important step in resolving these issues is to define the roles of the Arf GEFs at the molecular level through the identification of interacting partners, elucidation of membrane localization mechanisms, and analysis of Arf GEF mutants in vivo.

Identification of Interacting Partners of the Golgi-Localized Arf GEFs in Yeast and Mammalian Cells
Smirnova, Park, Niu, Phillips
The Arf GEFs of the Golgi-localized Gea/GBF and Sec7/BIG subfamilies are large multidomain proteins. A major goal of our laboratory is to understand the functions of the different domains of the Gea/GBF and Sec7/BIG Arf GEFs. Given that the GEFs are soluble proteins that must be targeted to membranes, identification of their membrane-targeting signals and partners is an important step in understanding their function. Within each subfamily, regions upstream and downstream of the Sec7 catalytic domain are conserved from yeast to humans. However, the functions of these homology regions remain unknown. We are carrying out two-hybrid screens with the Nterminal and C-terminal regions of the Arf GEFs of the Gea/GBF and Sec7/BIG sub-families. We carried out a two-hybrid screen using the Sec7 domain of the mammalian homolog of Sec7p, BIG2, and found a number of partners, two of which are human PI4P 5 kinase b and a putative lipase. PI4P 5 kinase b is an extremely interesting potential partner in that it has been identified as an effector of Arf. We carried out a reverse two-hybrid screen to determine the region of the BIG2 Sec7 domain that interacts with PI4P 5 kinase. Interestingly, we found that residues in the C-terminal region of the Sec7 domain abolish interaction with PI4P 5 kinase. We have narrowed down the interaction domain of PI4P 5 kinase to a 50-amino acid region in the Cterminal portion of the catalytic domain just upstream of the activation loop. The activation loop contains the substrate binding site and is both necessary and sufficient to determine intracellelular localization of the kinase. We are also mapping the region of Arf interaction in PI4P 5 kinase. Interestingly, it appears to be located in a different domain of the protein than the BIG2 Sec7 domain–interacting region. It is striking that we have identified three lipid-modifying enzymes as binding partners of different Sec7 domains. It appears that all three proteins bind to the C-terminal region of the Sec7 domain. We know that the region is adjacent to a PH domain in the ARNO subfamily of Arf GEFs, which mediates membrane localization through specific lipid interaction. Using the Gea2p N-terminal region, we have identified Mss4p as an interacting partner. Mss4p is the sole PI4P 5 kinase in yeast. It is intriguing that we identified a human homolog of Mss4p in our two-hybrid screen with the BIG2 Sec7 domain.

We have identified a five-span trans-membrane domain protein, Msg1p, as a potential membrane receptor for the Gea1p and Gea2p proteins. Msg1p was originally identified as a multicopy suppressor of the temperature-sensitive growth defect of the gea1-6 mutant, and it suppresses the cold-sensitive growth defect of arf1D when overexpressed. The gea1-6 mutant has two point mutations downstream of the Sec7 domain in the C-terminal region of the protein. Msg1p is highly conserved from yeast to humans, and both the yeast and human proteins are localized to the Golgi. In yeast, Msg1p co-localizes with Gea2p and with early Golgi markers. Msg1p interacts with the C-terminal region of both Gea1p and Gea2p by two-hybrid analysis, but not with the C-terminus of gea1-6. The results support the idea that Msg1p is a Golgi-localized membrane receptor for the Gea proteins.

Morphology of the Yeast Secretory Pathway
Park, Rainey; in collaboration with Hartnell, Rambourg
The Arf GEFs are important regulators of organelle structure and protein trafficking in both yeast and mammalian cells. Yeast organelles are seemingly differ dramatically in structure and organization from those of mammalian cells, yet most of the proteins involved in organelle structure and trafficking, including the Arf GEFs, are highly conserved. Current work is aimed at determining Golgi structure in yeast by using both live imaging and electron microscopy as a first step in identifying the structural features common to yeast and mammalian organelles. Use of both the yeast and mammalian systems will allow us to determine which aspects of Arf GEF function are fundamental to all eukaryotic organisms and which are unique to their specific system.

The molecular details of transport through the secretory pathway in the yeast Saccharomyces cerevisiae have been elegantly worked out. However, technical limitations have prevented the clear specification of the spatial organization of these molecular events. Organelles of the secretory pathway, in particular the Golgi apparatus, are difficult to image at both the light and electron microscopy levels. The difficulty results from the small size of yeast cells and the structure of the yeast Golgi apparatus, which is composed of approximately 30 dispersed elements scattered throughout the cytoplasm. The size of a yeast cell is only an order of magnitude above the wave length of visible light, thereby limiting the resolution of intracellular structures. It is therefore necessary for any structural analysis in yeast to include electron microscopy. We are collaborating with Lisa Hartnell to perform immuno-electron microscopy of the yeast secretory pathway. Previous work in collaboration with Alain Rambourg indicated that the large, ring-like structures that accumulate in the gea1-4 mutant are made up of tubular networks. Tubular networks of a similar type are also found in wild-type cells, although the structures are generally smaller. We were able to visualize the Och1-HA–containing structures in the gea1-4 mutant and found that they were very similar to those visualized by Rambourg’s reduced osmium technique. We were also able to obtain a good level of signal for the wild-type strain despite the fact that the Och1-HA–containing structures are smaller. We are currently carrying out co-localization experiments with early and late Golgi markers to determine whether they are present on completely separate Golgi elements or whether they can be present in different areas of the same continuous structure. We are also carrying out live imaging of secretory pathway proteins by using GFP-tagged markers.